search

mwhittaker / deeprl_project

Deep RL Final Project
1 stars 1 forks source link

Find the RL algorithm that will be improved by encoded states #1

Closed vlad17 closed 6 years ago

vlad17 commented 6 years ago

Find 2-3 candidates (W/ existing good 1GPU impls), pro/con, choose best, make running pong example.

vlad17 commented 6 years ago

Consider training speed too. ACKTR/CPO/DDPG might be faster real time training than dqn

SamvitJ commented 6 years ago

PPO baseline here: https://github.com/mwhittaker/deeprl_project/pull/11